SAMPL7: Host-guest binding prediction by molecular dynamics and quantum mechanics.

Area of Science:

• Computational chemistry
• Chemical physics
• Molecular modeling

Background:

• The Statistical Assessment of Modeling of Proteins and Ligands (SAMPL) challenges benchmark computational methods against experimental data.
• Accurate prediction of host-guest binding energies is crucial for drug discovery and materials science.

Purpose of the Study:

• To evaluate various computational approaches for calculating binding free energies of Octa Acid (OA) and exo-Octa Acid (exoOA) systems.
• To identify key factors influencing the accuracy of molecular mechanics combined with Poisson-Boltzmann or generalized Born surface area solvation (MMPBSA/MMGBSA) and quantum mechanics (QM) methods.

Main Methods:

• Molecular docking was used for initial pose generation, followed by molecular dynamics (MD) simulations.
• Binding free energies were computed using MMPBSA/MMGBSA and various density functional theory (DFT) approaches.
• Analysis included the impact of solvation models, partial charges, and dispersion corrections (e.g., DFT-D3).

Main Results:

• MMPBSA/MMGBSA methods showed improved correlation with experimental data when combined with empirical corrections.
• DFT calculations, particularly B2PLYP-D3 with larger basis sets, provided valuable insights, though B3PW91-D3 was better for geometry optimization.
• Accurate prediction necessitates careful consideration of dispersion effects and solvation models.

Conclusions:

• The study highlights the importance of incorporating dispersion corrections and appropriate solvation models for accurate binding energy predictions.
• Empirical corrections can enhance the performance of MMPBSA/MMGBSA methods.
• The findings contribute to the ongoing refinement of computational tools for molecular interaction studies.